“If today was perfect, there would be no need for tomorrow”. So what if perfection was just around the corner? What if that perfection superseded all human effort and all creativity? Would you continue to pursue perfection?

Around the second half of 2022, right on the heels of COVID 19, the landscape of sharing photos online changed. The combined release of Midjourney V3 in July and ChatGPT in December of the same year, marked a turning point in the way we think about images we see online.

As I’m writing this in late August of 2025, about 14.3 trillion photos have been taken since the 1800s (Photutorial 2025).
In 2024, about 15 billion images have been generated using various image generation tools (Everypixel 2024).
It took AI roughly 1.5 years to generate those 15 billion images. It took photographers worldwide 147,5 years longer to snap the same amount.
AI-generated images online will probably outnumber photos around 2040 if we extrapolate those numbers. But the big question is, how many photos people take actually end up online?

Seeking external validation

In my book Musings of a Landscape Photographer, I’ve written that not everyone is a photographer or artist for themselves. External motivational factors such as fame, money or securing social status and showing off the ability to travel the world are reasons for people sharing their images on social media. Other reasons include keeping up with peers or sharing memories with friends and family (Menon 2022).

This attention seeking behaviour is natural, as we need to secure a place in the social hierarchy of the tribe we live in. But Instagram isn’t a tribe, is it? And yet, I hear from my peers that there’s “a strong feeling of community” in the photography space online. To me, it has all the ingredients that lead to mental health issues.

The numbers are staggering. About 29% of teens feel pressure to post content that will get lots of likes and comments (Pew Research 2022) and half of adolescents using social media feels pressure to gain attention online (Dumas, et al 2023). In some parts of the world, kids and teens are even worse off. 87% of children aged 8 to 17 living in the UK say there’s pressure to be popular on social media at least some of the time (OFcom 2024).

“The algorithm”

I think it’s Meta with Facebook and Instagram who perfected how the purported black magic of the algorithm keeps us hooked. It’s not just teenage girls with a wrought self-image due to social media (The Social Dilemma 2020), but us photographers too who get sucked into the show off and comparison. Photographers left and right burn out as a result. I’ve now had a depression for about a year fueled by an existential crisis. Other well known landscape photographers (whom I shal not name) can’t keep up but feel the need they must and burn out.

There are countless articles written and videos circling on gaming the system and beating the algorithm. You can even buy fake followers online and have images be liked and commented on by AI to make you appear popular online for a monthly subscription (I would link to such services here, but I don’t want you to go down that rabbit hole).

Fake users

With your favorite social medium already infested with bots, we arrive at another big problem with social media.

The Ghost Data study in 2019 estimated the total number of bots on Instagram to be 95 million; which was about 9.5% of the total users at the time (Ghost 2019).
Meta themselves reported about 3% of the total accounts were dubbed “fake” in 2024 and 2025 (Meta 2025, sec.gov 2024).*
9-15% of bots on what used to be Twitter in 2017 were found to be bots (Varol et al. 2017). 5th Column AI analyzed that Elon Musk was right about his estimation of bots on Twitter pre-acquisition: 64% of users in 2024 was probably not human (Robinson 2024).

*Why 3% is a huge number

Computers can generate content non-stop, 24 hours a day, 365 days a year, at a rate that ‘s constantly increasing. Every fake account can put out content faster and faster and write nonsensical comments on your photos virtually simultaneously without being detected. Now imagine the entire population of Iran (92 million people) being robots constantly filling your feed with anything that’s not a photo. That’s just the number of bots on Facebook alone. You could assume 92 million is a conservative number as it’s estimated by Meta themselves. Here’s an estimation of my IG account from a source that’s not trying to convince shareholders.

“Damn kids these days” | Yet the world still changes

Media, technology and the world all change. Perhaps faster than we can tolerate. But shaking your fist angrily at the computer doesn’t help anyone. And AI definitely isn’t the first ‘tool’ that came along that changed photography. Color film photography made the art of black and white photography all but a quirky hobby for weirdos. And slide film allowed more saturated colors sparking controversy in purists in the age of analog even before the onset of digital cameras. And we’re not even talking about edits made in the dark room of which Ansel Adams himself was a big proponent. Adams is often attributed to internet one-liners such as “thou shalt not edit photos”, but the godfather of landscape photography himself surely would have found his way around in Photoshop, would he have lived as long.

At the start if the digital age, low megapixel CCD cameras came first. Those grainy digital images from the late 90’s were all shot on those. With their resolution far under performing compared to all film negatives and slides, digital photography was easily dismissed as fake, doctored or even unethical. With CCD sensors still being favored in cooled astrophotography setups, as they developed at their own pace, the onset and prevalence of CMOS digital camera sensors allowed for a higher dynamic range and eventually much higher resolutions than their CCD counterparts.

Photo editing software was, and often still is the next target of calling out “fake”. Photoshop in particular got a negative connotation for everything photography related. The term is often synonymous for the clone-stamp tool found within it, where anything from hot pixels to entire buildings can be “cloned” out to achieve a more pleasing result.

Photomatix, another app from the early 2000’s helped photographers to achieve the look of high dynamic range (HDR) images on a not-so HDR (SDR) screen. By mapping 32-bit tones onto what eventually became 8-bit files, we could choose how bright or dark any area in our photos was. Photomatix was a crude tool though. It led to unsightly halos and artifacts that gave the tool of choice away almost immediately. A few more of those came into being before Tony Kuyper and others gave us tools to make luminosity masking a breeze. Now, with actual HDR becoming more commonplace, we see the advent of true HDR photo-editing techniques. And I’m working on a tutorial to teach you how to do this properly in landscape photography. Photomatix was so popular, that around 2009, when I started to take landscape photography seriously, everybody used it.

Drone photography is another avenue that received a lot of scrutiny. It is supposed to be bad for the environment and wildlife. While I agree that drones can be a nuisance for both people and wildlife, it’s entirely up for debate if it has a negative impact on the environment. It certainly depends on the discretion of the pilot, local legislation and the area in which the drone is operated. You can perfectly fly at 120 meters without disturbing any birds flying much lower to get your cool abstract images. On the other hand, I’ve seen drone pilots taking off from right behind a sign saying “NO DRONES” at Fjaðrárgljúfur in Iceland…

Relatively new are tools like Luminar, in which the generated sunrays in particular made headlines. Every time a sunray in any photo was posted post 2017, “Luminar” was the word you’d read in the comments. As if calling out the tool immediately invalidated the image. That’s where we are now, again. In 2025, we see the same thing happen with another buzzword in the business of photography: AI.

AI is different / just another evolution of photography

I’m in the red corner: I think AI will eventually destroy humanity.

Whoa, that escalated quickly. Let me slow down, add some context and explain why I think AI is different. First, let’s separate generating images and AI-assisted editing.

Like (matte-)painting, generated content isn’t photography. It’s just not the same as having shot an image that you edit in any way. Even Luminar’s sunrays started out with a photo. It’s an effect that added in something that wasn’t there in the field. Unlike blending two images; while both of them still shot by you, the effect is entirely generated with the help of software. Where AI generated images differ is that no photography is required to create an image that may or may not look like your own photography. This is important, because it certainly can create images that look indistinguishable from photography. Even a year ago, when I started an experiment that would drive away most of my following on Instagram and got me banned from 500px, generated images were not identified as not being photos.

The experiment

The experiment involved me posting images on Instagram and 500px that were entirely generated by Midjourney, while I fed it my own photos to create images that looked like it. Sometimes using /Blend, where two or more of my own photos were mixed together to create a new image in the same vein. The experiment ran for months. Images of forests became progressively wilder, as I was thinking “how come nobody sees this”? It wasn’t until the first commenter noticed and asked if it was AI that I admited to it not being a photo I took and edited myself. On 500px, a dying photography sharing platform since it was bought by a Chinese conglomerate in 2018, I ran the experiment as well. Until I added AI in the description, no one noticed or gave a damn. Two or three images later and I couldn’t log in. My account was deleted and me logging in was out of the question. Somehow I violated the terms of service and was banned from the platform without warning. I tried fighting it by sending an email, but to no avail. My blogs that I’ve written for 500px when it was still a small San Fransisco based company with great people as editors are still there, linking to images that are not found and pointing to dead links. I’m not sad the profile is gone, because I’ve grown tired of using social media anyway, but it is kind of weird that I was able to upload a whole bunch of AI slop until the point where I said that it was AI slop.

AI again. Notice the weird noise? That’s the stable diffusion pattern hard at work. But mind you, these are images generated in 2023!

Now it’s the other way around

We’ve now arrived at the reason why I’ve decided to write a blog post about this. Following my previous blog about AI, where I wrote that generating images is fun to me, it’s time to point out that a change is coming.

A few months ago I noticed, rather frustratingly, that a growing number of people now assume my photos are AI first. With the experiment from above, that seems only natural. But I’ve started to share my images on Reddit after the experiment, to get sense of how people would react to my photos in a place that’s not familiar with my work. Since Reddit is not afflicted by algorithms hiding or shadow-banning your photos for people who actually follow you, while having a respectable user base, I thought that Reddit is the place for sharing photos today. Apart from the few “beautiful photo” comments, this is what’s brewing:

Yeah, the footsteps look real perfect. As does the flock of seagulls on the left. In my eyes, perfection is what we strive for, but never achieve in photography. Nor should we. “If today was perfect, there would be no need for tomorrow” was on my wall for years.

*Only if it’s real though. If it’s not, I hate it.*

*Never post raw files. Ever. Also, it’s raw. No capital letters. It’s not an abbreviation. And English is not my first language.*

*Thanks. I don’t worry, but I do take note and get a knot in my stomach that leads to articles like these.*

Soon, you as a photographer will have to defend the fact that your work is not created by a computer. To me, this is insane. But wait, I can back it up.

Instagram’s “Made with AI” label created false positives (PetaPixel 2024)
YouTube wants your ID (Venus Theory 2025), shown below.
Photography platforms allow AI generated content in specific categories (One EyeLand)
Photo contests won with AI-generated images (Australian Photography 2023, The Guardian 2023)
…and vice-versa: AI-contest won by photographer and gets disqualified (ARTnews 2024)

It’s not just YouTube suffering the problem of user accounts and videos being entirely generated with AI. It’s everywhere. Which leads me to my next point.

What’s the point?

If all of these platforms see progressively more bots and computer-generated content, why do we bother visiting those platforms anymore? And why do we bother sharing our work on the same platforms?

The most important point that my book Musings addresses comes up yet again. If no-one online cares: are you still as excited about photography?

My personal journey (because everyone is on a journey nowadays) led me to the opposite of what you might assume about me after reading this rant about technology escaping my control: since the whole AI thing started, I stopped looking at social media and other people’s “photos”. I took a cold hard look at what inspires me and why I took up photography in the first place. It turns out that I like crossbreeding fantasy with reality in way that’s pretty unique according to Cody Schultz and Ben Horne discussing my odd sources of inspiration.

Sometimes that means creating photos that resemble scenes from a movie. Sometimes that means I recreate Magic the Gathering-esque paintings in post from raw-files that have people scratching their heads. Other times I use music to shape the thoughts whirling through my mind. Most of the time though, I don’t even share my images. Maybe they’ll pop up in my newsletter eventually.

Photography to me is like building a sand castle close to sea: photos are as fleeting as the moment you capture them, only slightly less ephemeral as life itself.

Why AI will challenge what it’s like to feel alive and to be human

If you’ve not read the paper AI 2027, I urge (not recommend but urge) you to do so. If you don’t want to read through it, here’s a YouTube video of a channel dedicated to it:

With human-machine augmentation on the rise, I think creativity and consequently, the will to share and see images online will lose it’s meaning well before 2030. Today, less than 100 people have an invasive, medical-grade brain-computer interface like and including Neuralink (Reuters 2025, Endovascular Today 2024, Onward Medical 2025, Neurology 2023).

Looking at other people’s promt results is as fascinating as watching paint dry, although having your wall done immaculately is immensely satisfying. ChatGPT’s estimation when we reach the first million people with something like Neuralink? Somewhere between 2041 and 2053.

I think that’s the 100th monkey; the critical mass people need to embrace it as normal. We’ll probably use that stuff even when we don’t need it. To be faster and more creative than the others. Next post will be less doom and gloom, because even though the future looks bleak we still have now. And to be at peace with the here and now is what will make you a better photographer.

The AI Photo Landscape | What happens in 2027?